Hi, Can I train DeepFace or VGG16 + Siamese model to locate face landmarks? I want to reach something like [this](https://camo.githubusercontent.com/37bc4cdc15e60f3155228c0660a2aab9c5ac0303/68747470733a2f2f6f6f6f2e306f302e6f6f6f2f323031372f30362f32392f353935346338313139346436392e706e67). I am planning to use [CelebA ](http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html)dataset. Thanks.
